CHAP T E R 8

Nanomaterials: Synthesis and

Characterization

8.1 SYNTHESIS OF NANOSCALE

MATERIALS AND STRUCTURES
There are many ways of making nanoscale materials and structures.
Figure 8.1 orders them in the same matrix as that of Chapter 6. It
gives an overview, showing where each fits. We start our discussion
at the upper left: processes to make discrete, nanoscale clusters.

Methods for Making 0-D Nanomaterials

Nanoclusters are made by either gas-phase or liquid-phase proc-
esses. We start with those using the gas phase, the commonest of
which are inert-gas condensation and inert-gas expansion. Liquid
phase processes use surface forces to create nanoscale particles and
structures. There are broad types of these processes: ultrasonic dis-
persion, sol-gel methods, and methods relying on self-assembly.

Inert-gas condensation
In inert-gas condensation, an inorganic material is vaporized inside a
vacuum chamber into which an inert gas (typically argon or helium)
is periodically admitted (see Figure 8.2). The source of vapor can be
an evaporation boat, a sputtering target, or a laser-ablation target.
Once the atoms boil off, they quickly lose their energy by colliding
with the inert gas. The vapor cools rapidly and supersaturates to
form nanoparticles with sizes in the range 2–100 nm that collect
on a finger cooled by liquid nitrogen. The particles are harvested by
scraping them off the finger and are collected, still protected by the
inert gas, for further processing. Alloy particles are made using dual
sources, as shown in Figure 8.2. The main problem with this method
is that as the particles form, they tend to cluster and increase their
Nanomaterials, Nanotechnologies and Design
Copyright 2009 Elsevier Ltd. All rights of reproduction in any form reserved. 257
258 C H A P T E R 8 Nanomaterials: Synthesis and Characterization

Classes

Class 1 Class 2 Class 3

Discrete nano- Surface nano- Bulk nano-
objects featured materials structured materials
All 3 dimensions
on nanoscale

Inert gas condensation Physical or chemical Equiangle extrusion

Cryomilling
0-D

Evaporation vapor deposition Consolidation of

Colloidal methods (PVD or CVD) nanoparticles by sintering
Dimensionality

on nanoscale
2 dimensions

Directional growth from Incorporation of nanotubes and

1-D

catalyst dots Lithographic methods rods into polymer or

Templating metal matrices
on nanoscale
1 dimension

Beating (gold foil) Electrodeposition

Electrodeposition
Rotating shutter PVD and CVD
2-D

Physical vapor deposition (PVD) Cyclic electrodeposition

PVD, CVD Chemical vapor deposition (CVD)
Self-assembled films

FIGURE 8.1
The matrix of processes used to make
size, defeating the original objective. To avoid this, the processing
nanomaterials.
parameters have to be carefully controlled.

Cold finger
Inert-gas (free-jet) expansion
Scraper
In inert-gas or free-jet expansion, evaporated atoms are carried by a
high-pressure helium gas stream that is expanded from a nozzle
into a low-pressure chamber at supersonic velocities (see Figure
Clusters
8.3). The adiabatic expansion of the gas leads to sudden cooling,
causing the evaporated atoms to form clusters a few nanometers
Inert gas
Evaporation in diameter. As in the case of inert-gas condensation, the agglom-
sources
eration of nanoparticles is a problem, requiring careful control of
evaporation rate and inert gas flow; it is these that determine the
particle size and distribution.

Vacuum Sonochemical processing

pump
Collection In sonochemical processing, ultrasound is used to nucleate a chemical
tray reaction. Ultrasound spans the frequencies in the range 15 kHz to
FIGURE 8.2 1 GHz. A magnetostrictive or piezoelectric transducer (or “horn”)
Inert gas condensation. The nanoclusters that form is used to generate ultrasonic waves with a wavelength of 1–10,000
on the cold finger are harvested by the scraper microns in a liquid-filled reaction vessel (see Figure 8.4). These
and collector and subsequently consolidated to are not molecular dimensions, so there is no direct coupling of
make products. the acoustic field with the chemical species. The reaction comes
 Synthesis of Nanoscale Materials and Structures 259

about because of cavitation. The tensile part of the wave is intense Clusters

enough to pull the liquid apart and form a tiny cavity. The compres-
sion part of the wave then compresses it, but before it does, some Helium
gas
reactants vaporize inside it. The next tensile wave re-expands the
bubble, which oscillates in volume at the frequency of the sound Evaporation
Collection
sources
tray
waves, pumping it up as more vapor enters during the expansion Vacuum Vacuum
pump pump
part of the cycle. When the bubble reaches a critical size, it collapses.
The collapse is adiabatic because the very fast collapse rate leaves no FIGURE 8.3
time for heat flow, generating a tiny, localized hot spot. The tem- Cluster formation by vapor phase expansion from
peratures are very high (as high as 5000°C, near that of the surface an oven source.
temperature of the sun) and so too are the pressures (around 2000
atmospheres, roughly those at the bottom of deep oceans), trigger-
Ultrasonic
ing reactions that create a nanoparticle within the spot. The size horn
of the spot determines the size of the resulting particles. By using
organometallic precursors, ceramic and metallic particles as small
as 2 nm can be produced. The technique can be used to produce a
large volume of material for industrial applications. Reaction
vessel

Sol-gel deposition Pressure

Ultrafine particles, nanothickness films, and nanoporous mem- waves
branes can be made by sol-gel processing (see Figure 8.5). The start-
Reacting
ing point is a solution of precursors in an appropriate solvent. hotspots
The precursors are usually inorganic metal salts or metal-organic
compounds such as alkoxides—metal ions with an organic ligand Nano-
particles
such as Ti (OC4H9)4. The precursor is subjected to a polymerization
reaction to form a colloidal suspension, or “sol,” of discrete, finely
dispersed particles kept in suspension by adding a surfactant. The
suspension can be treated to extract the particles for further process-
ing, or it can be cast or spin-coated onto a substrate. There it is Tension
pressure

converted to a gel by chemical treatment to disable the surfactant to

Sound

create an extended network of connected particles throughout the

solution, making a kind of superpolymer, one enormous molecule
Compression
in the form of an open 3-D (or, on a surface, a 2-D) network—the
“gel.” Evaporation of the solvent then leaves a dense or nanoporous
film. Sol-gel methods are the basis for a wide variety of materials,
diameter

including paints, ceramics, cosmetics, detergents, and cells.

Cavity

Cavity growth

Molecular self-assembly
Molecular self-assembly methods rely on the self-organization of Cavity collapse
organic molecules. The most obvious is that of crystallization: Time
Cool a saturated solution of sugar or salt and the molecules self- FIGURE 8.4
assemble into crystals. Nature uses self-assembly in infinitely subtler Sonochemical processing. The ultrasound induces
ways; indeed, the whole of the natural world is self-assembled. cavitation. Cavity collapse causes the reagents to
Perhaps the most remarkable of all is the self-assembly (and self- react.
260 C H A P T E R 8 Nanomaterials: Synthesis and Characterization

disassembly) of DNA that passes genetic information from cell to

cell, instructing them how to self-assemble. Surely we, as designers
and engineers, could use this path, too.
We are only just beginning to exploit its potential. The idea is to
Metal-alkoxide
create the conditions under which atoms or molecules will self-
solution assemble into useful structures, driven by the minimization of
Hydrolyze, their energy. The great advantage of self-assembly is that the system
polymerize converges to a specific configuration without the need for further
control. Typically, the aggregates formed by self-assembly tend to
be bonded by relatively weak bonds with binding energies only a
Precipitation few times larger than kT, the thermal energy per atom.1 The self-
assembling molecules form micelles. Micelles are aggregates of
amphiphilic molecules—molecules with one end that is soluble
Sol Discrete in water and the other end that rejects it. These aggregates form
nanoparticles spontaneously at a size that depends on the concentration of the
Gelling
agent
amphiphilic molecules in solution. The center of the micelles acts
as a chamber for chemical reactions and thus dictates the size of the
nanoparticles created (see Figure 8.6).
Solvent
evaporation Self-assembly of 2-D nanofilms is made possible by using the
wonderfully named Langmuir-Blodgett technique (see Figure 8.7).
A monolayer of a fatty acid is created on the surface of water into
which a substrate has been placed. Fatty acid molecules have a
Gel Nanoporous
membrane
hydrophilic part and a hydrophobic part. The polar part locks into
the water, whereas the nonpolar part rejects it. The fatty acid self-
FIGURE 8.5
spreads across the water surface as a monolayer; any other arrange-
Sol-gel processing can be used to make
ment would bury one or another end of molecules where they did
nanostructured layers and coatings as well as
nanoporous membranes.
not want to be. The substrate is slowly withdrawn from the water.
The monolayer sticks to it and is transferred from the water surface
to the substrate. The process allows the fabrication of single mon-
olayers of materials as well as thicker films by repeatedly dipping
and withdrawal.

Methods for Making 1-D and 2-D Nanomaterials

The production route for 1-D rodlike nanomaterials by liquid-phase
methods is similar to that for the production of nanoparticles. Self
assembly methods use the highly anisotropic bonding nature of
asymmetric molecules to cause them to self-assemble into tubes
rather than spheres, forming cylindrical micelles. The amphiphilic
molecules are then removed with an appropriate solvent or by cal-
cining to obtain individual nanowires.

1
Here k is Boltzmann’s constant and T is the absolute temperature.
 Synthesis of Nanoscale Materials and Structures 261

CVD methods have been adapted to make 1-D nanotubes and Hydrophobic
tails
nanowires. Catalyst nanoparticles are used to promote nuclea-
tion. To make carbon nanotubes, for example, a combination of Hydrophilic
carboncontaining gasses, such as methane (CH4) and/or carbon heads
monoxide (CO2), are reacted in the presence of Iron (Fe), Cobalt
(Co), and Nickel (Ni) catalysts at 1100°C. Decomposition of the
gasses releases free carbon atoms that condense on the substrate
with its array of catalyst particles, from which the carbon nanotubes
grow. In an alternative process, an arc is generated between two elec-
trodes, one of which is carbon. The arc creates high temperatures, Water
droplet
causing the vaporization of the carbon electrodes into a plasma.
The arc is typically operated in a gas environment, such as nitrogen
or helium, and is generated by an electrical current passing through
the electrodes, which causes the ionization of gas atoms. The ion
beam produced is directed to a substrate carrying the Co, Fe, or Ni
catalyst particles from which the carbon nanotubes grow.
Nanowires of other materials such as silicon (Si) or germanium (Ge) FIGURE 8.6
are grown by vapor-liquid-solid (VLS) methods. Typically a catalyst Self-assembly of nanoclusters. The enclosed
seed (gold, for example) is deposited on a substrate. Subsequently a droplet can be spherical or cylindrical. Reaction
vapor phase with the desired composition is brought in contact with within it generates a nanoparticle or nanorod.
the seed at a controlled temperature. The vapor diffuses into the cat-
alyst, changing its composition and lowering its melting point until Substrate
it melts. The liquid surface has a large coefficient of accommoda- Monolayer
tion and therefore acts as a preferred site for absorption of the gas of fatty acid
vapor. The liquid becomes supersaturated and a solid nanowire or
“whisker” then grows from it with a diameter equal to the diameter
of the catalyst seed. Nanowires with a diameter around 10 nm and a
length of 1 micron can be grown in this way.
We have just encountered two ways of making layers of nano thick-
ness: sol-gel methods and the method of Langmuir-Blodgett films.
There are others, described in the following subsections. FIGURE 8.7
Self-assembly of 2-D nanofilms. Langmuir-
Blodgett film formation and extraction.
Foil beating
Gilding is the art of applying gold leaf to a surface. It was known to Hammer
the ancients, is mentioned in the Old Testament, and was described Gold foil
by Herodotus in 420 BC. Gold is an expensive material; there are Anvil
obvious incentives to use as little of it as possible. Gold leaf is made
by beating the metal to a uniformly thin foil, roughly 100 nm thick
(see Figure 8.8). It is applied by smoothing it onto the adhesive-
coated surface of the object to be gilded.
There are less labor-intensive ways to make thin films. Electrodeposi- FIGURE 8.8
tion is one, and it is one able to create nanoscale structures as well. Beating of gold leaf.
262 C H A P T E R 8 Nanomaterials: Synthesis and Characterization

Electrolyte Electrodeposition
Substrate - + Electrodeposition (see Figure 8.9) is a long-established way to deposit
metal layers on a conducting substrate. Ions in solution, some-
M+
times replenished from an anode, are deposited onto the nega-
tively charged cathode, carrying charge at a rate that is measured as
a current in the external circuit. The process is relatively cheap and
fast and allows complex shapes. The layer thickness simply depends
on the current density and the time for which the current flows. The
deposit can be detached if the substrate is chosen to be soluble by
Nanocrystalline Anode dissolving it away.
deposit
Metal The challenge is to control the structure. Most electrodeposits
ions
grow in a columnar manner: Metal crystals nucleate on the bare
substrate when the current is switched on and grow outward as
interlocking pillars, like a miniature Giant’s Causeway. To make
Volts nanoparticles or layers, it is necessary is to stop each crystal’s
growth while it is still tiny and to nucleate more. A combina-
Time tion of tricks is used to do this. Pulsing the voltage, as sug-
FIGURE 8.9 gested by the graph in Figure 8.9, causes little bursts of crystal
Pulsed electrodeposition. The pulsed potential growth. Adding growth inhibitors that condense on the crystal
nucleates many nanoscale crystals that build to surfaces during the “off” phase of the pulse discourages their
form a layer up to 5 mm thick. continued growth during the next “on” phase. These methods,
combined with a high current density, nucleate many crystals
but allow little time for each to grow, giving a nanostructured
deposit.
The technique can yield porosity-free nanocrystalline deposits as
Substrate thick as 5 mm that require no further processing.
Deposit

Physical vapor deposition (PVD)

In PVD plating, a thin layer of a material, usually a metal, is depos-
Evaporated ited from a vapor onto the object to be coated (see Figure 8.10).
atoms The vapor is created in a vacuum chamber by direct heating or
Heated
electron beam heating of the metal, from which it condenses onto
evaporant the cold substrate, much like steam from a hot bath condensing
Heater
on a bathroom mirror. In PVD plating there is no potential differ-
ence between bath and work piece. In PVD ion plating the vapor
is ionized and accelerated by an electric field (the work piece is
Vacuum the cathode, and the metallizing source material is the anode).
FIGURE 8.10 In PVD sputtering, argon ions are accelerated by the electric field
Physical vapor deposition (PVD). Material, onto a metal target, ejecting ions onto the component surface. By
evaporated by heating, by ion bombardment, or by introducing a reactive gas, compounds can be formed. (Sputtering
laser ablation, is deposited on a substrate target. titanium in an atmosphere of nitrogen, for instance, gives a coating
The deposited layer can be of nano thickness. of hard TiN.)
 Synthesis of Nanoscale Materials and Structures 263

Almost any metal or compound that doesn’t decompose chemi-

cally can be sputtered, making this a very flexible (though expen- Induction or
sive) process. Targets can be changed during the process, allowing radiation
multilayers to be built up. These multilayers can have remarkable heating
mechanical and electronic properties. Reacting
gasses
Chemical vapor deposition (CVD) Components
In CVD processing, a reactant gas mixture is brought into contact Coating
with the surfaces to be coated, where it decomposes, depositing a
dense pure layer of a metal or compound (see Figure 8.11). The Reactor
deposit can be formed by a reaction between precursor gases in the
vapor phase or by a reaction between a vapor and the surface of
the substrate itself. A difficulty with the process is that it frequently
requires high temperatures, 800°C or more.
Gases Gases
In a variation of conventional CVD called moderate temperature CVD in out
(MTCVD), metal organic precursors are used (hence its other name, FIGURE 8.11
metal organic CVD, or MOCVD). As they decompose at relatively Chemical vapor deposition (CVD). Reaction of
lower temperature, the reaction temperature is typically around gases in the reaction chamber causes deposition
500°C. If the chemical reactions in the vapor phase are activated by of the reaction product as a layer that can be of
the creation of a plasma in the gas phase or by shining a laser beam nano thickness.
into the gas mixture, deposition at an even lower temperature, a
little above room temperature, becomes possible. These techniques
are called plasma-assisted (or plasma-enhanced) CVD (PACVD or
PECVD) and laser CVD (LCVD).
Coatings formed with PCVD methods are typically nanocrystalline
or amorphous because their formation is no longer dependent on
equilibrium thermodynamic constraints.

Methods for Making 3-D Nanomaterials

Many bulk materials that we use today derive their properties from
internal structure at the nanoscale. High-strength, low-alloy (HSLA)
steels get their strength from nano-dispersed carbides. All the alloys
of aerospace—based on aluminum, magnesium, and titanium—
are strengthened by precipitates with dimensions in the 10–100 nm
range. The copper alloy wiring connectors in automobiles and
other electrical equipment retain their springiness, maintaining
good electrical contact, because of nanoparticles of CuBe. All of
these (and there are many others) are conventional alloys contain-
ing a dispersion created by heat treatment. The dispersion occupies
only a small fraction, typically 1–5%, of the volume. The matrix—
the steel, the aluminum, the copper—is not itself nanocrystalline.
264 C H A P T E R 8 Nanomaterials: Synthesis and Characterization

Scale

Macro
mm Top down

Equiangle extrusion
Rapid solidification

Micro Intermediate
mm

Mechanical alloying
Micromachining Bulk nanostructured
solid

Nano Bottom up
nm
Inert gas condensation
Sol-gel and self-assembly
MFA, 07

FIGURE 8.12
The top-down, intermediate, and bottom-up
Suppose you could make it so? It’s not easy. But materials scientists
approaches to making bulk nanostructured solids.
like a challenge, and they have risen to this one.
There are three directions of attack, sketched in Figure 8.12. The first
is to break down the structure of a bulk material, reducing the crystal
size to submicron or nanodimensions—the top-down approach. The
second—the intermediate approach—starts with standard micron-
scale particles, reducing their structure to the nanoscale by milling
techniques. The third is to build the solid up from the atomic scale
or from nanoclusters in a way that retains the scale of is structural
units; this is the bottom-up approach.

Top-Down Processes
Top-down processes are best described this way: Take the material
and drug, beat, and freeze it. Drug it by alloying. Beat it by severe
plastic deformation. Freeze it to stop the fine-scale structure, once
formed, from coarsening.

Rapid solidification
Materials would rather be crystals than glasses. If ordered, all atoms
sit at exactly the distance from their neighbors that best satisfies
 Synthesis of Nanoscale Materials and Structures 265

their interatomic bonds. Disorder disrupts this comfortable seating

arrangement, stretching some bonds and squeezing others. Liquids
are disordered because heat shakes the atoms so violently that they
Reservoir of
are sprung from their low-energy, crystalline arrangement. The molten alloy
melting point is the temperature at which this disruption occurs;
below it the thermal shaking is too weak to disrupt bonds, and the Jet of
molten alloy
liquid crystallizes.
Cooled at normal rates, most liquids solidify to give solids with Water-cooled
large crystals, or “grains.” The trick in making nanocrystalline or spinning drum
amorphous materials by casting them is to deprive the material of
the time or the means to transform from liquid to solid. Some mate-
rials, of which window glass is one, are easily duped into retaining
their glassy structure—their high viscosity when liquid slows the
rearrangement of the molecules to form crystals. Many polymers,
too, are “glassy” because their tangled molecules cannot reorganize Amorphous or
in any normal timeframe to form the crystal they would like to be. nanocrystalline
strip
Metals and ceramics, by contrast, crystallize at the drop of a hat. It
takes extreme measures to make them retain their liquidlike struc- FIGURE 8.13
ture or adopt a structure with exceedingly small grains. Melt spinning allows cooling rates up to 1 million
degrees Centigrade per second. This is enough
The first step is to mix in elements with different-sized atoms, each to freeze in the liquid structure in certain alloys,
preferring a different atomic spacing and crystal structure, making giving amorphous or nanocrystalline wires or
crystallization difficult. The second step is to cool quickly, leaving ribbons.
little or no time for crystallization. Early alloys required precipitous
cooling rates exceeding 1,000,000°C/sec, limiting the form to thin
wires and ribbons from which heat can be conducted quickly.
Figure 8.13 shows one way of achieving such cooling rates. A jet
of liquid alloy—one designed to be hard to crystallize—is squirted
onto a spinning, water-cooled copper drum. The process is called
melt spinning. The liquid layer cools fast enough to become amor- Laser
phous or, if not that, then nanocrystalline. The laminated cores of
Focused
many transformers and the read/write heads of magnetic tape and beam
disc recorders are made that way, exploiting the special magnetic
properties of the amorphous state. Newer bulk-amorphous metals Amorphous
layer
(BAMs) remain glassy even at relatively slow rates of cooling (10°C/
sec), allowing thick sections (up to 20 mm) to be cast.
Figure 8.14 shows a second way of cooling a liquid fast: Zap it with
a laser beam, scanning the beam fast enough that it melts only a
very thin surface layer. The layer is already stuck to the cold material FIGURE 8.14
beneath. Conduction of heat into this cold substrate is fast enough Laser surface hardening. The laser beam melts
to trap the amorphous structure. This laser surface melting is the way an exceedingly thin layer, which then cools so
amorphous silicon for cheap solar cells is made. The same tech- fast by conduction that it becomes amorphous or
nique is used to make nanocrystalline surface layers. nanocrystalline.
266 C H A P T E R 8 Nanomaterials: Synthesis and Characterization

Electroless There are other ways to make amorphous structures. One is to grab
nickel coating the atoms or molecules not from the melt but from solution. Electro-
deposition, already described (Figure 8.9), can be driven at such a
rate that the ions, dragged to the surface of the cathode (nega-
tive electrode), are hurled onto it so fast that they have no time
Ni ions Ni ions to rearrange and build crystals. More remarkably, the deposition
can be done, in one case, without any electric current. Electroless
nickel (really a nickel-phosphorous alloy) deposits spontaneously
onto almost anything plunged into a bath containing nickel ions
and a reducing agent, sodium hypophosphite (see Figure 8.15).
The deposit is uniform (unlike one made by electrodeposition)
and is exceptionally hard and resistant to wear and corrosion. And,
Sodium hypophosphite
solution
crucially, it is cheap.

FIGURE 8.15 Equiangle extrusion

The electroless nickel process. When a part with a
suitably activate surface is plunged into the bath, a
hard, amorphous, nickel-phosphorous layer plates
onto its surfaces.
Plunger
The second approach is that of extreme plastic deformation. Figure
8.16 shows equiangle extrusion—a way of generating extreme defor-
mation while suppressing fracture. A rod, forced through the die,
is savagely sheared and extruded, emerging as a rod with the same
diameter as that with which it started but with a much refined
structure. It can then be reinserted into the die and tortured further
until the structure is sufficiently refined. The process is now a stand-
ard one for making metals with grain sizes in the 100–500 nm
range.

Equal Intermediate Processes

channel
die
Milling and mechanical alloying
Equiangle extrusion leaves the composition unchanged. Milling,
Structure shown in Figure 8.17, combines extreme deformation with the
before
extrusion
ϕ forcible alloying of two materials that, normally, would not mix.
Particles (here, two metals, A and B) are spun in a high-energy ball
mill. The heavy steel or carbide balls are thrown against each other,
trapping and squashing the metal particles between them; the par-
ticles flatten, weld, and then break up. The process creates heavily
Severe deformed, mechanically alloyed particles with (if continued long
deformation enough) a nanoscale internal structure. It helps to fill the mill with
Structure after inert gas to promote welding and to cool it (cryomilling) to prevent
extrusion the structure coarsening. The particles are subsequently compacted
Extruded and sintered to make products.
material
FIGURE 8.16 Micromachining
Equiangle extrusion. Each pass shears and breaks There is another way to torture materials on a small scale. It is a
up the structure; repeated, the process reduces it familiar one: machining. When you look at a machined object,
to the nanoscale. you see the technical precision, the clean, engineered quality,
 Synthesis of Nanoscale Materials and Structures 267

that distinguishes machined products form those that are merely Rotating drum
molded or cast. What you don’t see is the swarf—the machined
shreds, chips, and tangles. These have suffered extreme deforma-
tion; their structure (like those of the last two processes) is refined
by it. There is an interesting scaling law here: The finer the cut, the
thinner the chip and the greater the shear it has suffered. Microma-
chining—machining on the scale of watch making, as in Figure
8.18—deforms the chips more than ordinary coarse machining.
Better yet is machining with a diamond tip with submicron radius.
Diamond nanomachining is designed for surface profiling, but the
swarf it produces can be useful too. As with particle milling, a way
must be found to consolidate the swarf into useful products. We Component A
get to that next. Component B

Steel or
Bottom-Up Processes ceramic
balls
Consolidation of nanoclusters and milled powders
Most of the processes described thus far do not make solid objects;
they make clusters, powders, or chips. The long-established way to
consolidate powder is to press in a die that has the desired form,
then heat to a temperature at which diffusional bonding takes place.
This powder-pressing and sintering route to manufacturing products
(see Figure 8.19) is widely used to make engine parts for cars and
Repeated
components for household appliances such as washing machines. extreme
The difficulty in using it to consolidate nanostructured particles is deformation
that sintering takes time, and at the sintering temperature, the struc- and cold welding
ture coarsens. It is not an easy problem to solve; during consolida- FIGURE 8.17
tion there will always be some coarsening. The question is how to Powder milling with mechanical alloying. The
minimize it. One way is to compact the powder in such a clean heavy steel or tungsten carbide balls trap, deform,
environment that the particles bond, even at room temperature, weld, and break up the powder particles, mixing
but that is seldom possible. An alternative is to sinter so fast that them so completely that they become alloys with a
there is little time for coarsening. Figure 8.20 shows one way to do nanoscale structure.
this—that of flash sintering. The powder (or swarf) is compressed in
a die through which a bank of capacitors is discharged. The blast of
heat, generated by the resistance of the packed powder, is enough
to create good bonding without leaving enough time for serious
coarsening.

Methods for Nanoprofiling

Often it is not a nanomaterial that is sought; it is nanofeatures
on the surface of something much bigger. These are created by
micromachining (cutting material away) or by microlithography
(putting material where you want it).
268 C H A P T E R 8 Nanomaterials: Synthesis and Characterization

Cutting Micromachining
tool
Intricate 2-D and 3-D patterning of materials is made possible by
Heavily techniques derived for those of conventional machining. Table 8.1
deformed
chip lists these with their approximate resolution limits. The current
resolution of micromachining (see Figure 8.21) and micro EDM
Work
piece (electro-discharge machining), the methods of watch makers,
is 3–5 microns. Machining with focused electron or laser beams
(see Figures 8.22 and 8.23) allows submicron resolution with high
Motor removal rates. Focused ion-beam (FIB) machining offers the great-
est resolution, with the ability to make features as small as 20 nm,
but it is very slow. In FIB machining a beam of gallium ions from a
liquid metal ion source is accelerated, filtered, and focused with elec-
tromagnetic lenses to give a spot size of 5–8 nm (see Figure 8.24).
The beam is tracked across the surface to be machined, contained
in a chamber under high vacuum. The high-energy ions blast atoms
from the surface, allowing simple cutting of slots and channels or
Collected
swarf the creation of more elaborate 3-D shapes. Secondary electrons are
FIGURE 8.18 emitted when the gallium ions displace the surface atoms. These
Micromachining. The machined chips are heavily can be used to image the surface, allowing observation and control
sheared, reducing the scale of their structure. of the process as it takes place. Dual-beam FIBs have an additional
electron gun that is used as an alternative way of imaging. The pre-
cision is extraordinary, but the process is very slow.
Pressure

Table 8.1 Micromachining Methods

Machining Materials Feature Positional Material
Method That Can Be Size (and Tolerance Removal
Machined Tolerance) Rate,
Microns3/
sec
Micromachining Metals, 10 microns 3 microns 10,000
polymers (2 microns)
Heated Micro Any 10 microns 3 microns 2,500,000
die electrodischarge conducting (3 microns)
machining (EDM) material
Powder
Electron beam Any 5 microns 1 micron 100,000
FIGURE 8.19 machining (EBM) conducting (submicron)
Pressure sintering, the standard way to material
consolidate powders. Heating is slow, allowing
Femto-second Any material 1 micron Submicron 13,000
time for the structure to coarsen.
laser machining (submicron)
(LBM)
Focused ion- Any material 0.2 microns 0.1 microns 0.5
beam machining (0.02
(FIB) microns)
 Synthesis of Nanoscale Materials and Structures 269

Upper Turret Lathe

punch
Vacuum Capacitor Guide
chamber bank plate
Motor

Motor Motor
Graphite
mould

Powder Motor

Lower
punch Work
piece

FIGURE 8.21
FIGURE 8.20
A micromachining workstation. The resolution limit
Electric discharge, or “flash,” sintering. The rapid
is about 5 microns.
consolidation leaves little time for structural
coarsening.

Laser
Vacuum chamber (10-5 mm Hg)
source
Laser
Electron beam
Electro-
gun Plasma
magnetic Laser cut
lens Electron slot
beam
Beam
deflector Machined
slot

Parent
metal
FIGURE 8.22
FIGURE 8.23
Electron beam machining. The resolution limit is
Machining using a femto-second laser. The
about 1 micron.
extremely short pulse time allows a machining
resolution around 1 micron.

Photolithography
Photolithography, the process used to manufacture computer chips,
is able to produce features smaller than 100 nm (see Figure 8.25).
An oxidized silicon wafer is coated with a 1µm thick photoresist
layer, as in Figure 8.25a. A beam of UV light is incident on a mask,
allowing light to pass through the gaps. The light is then passed
through a set of lenses to reduce the pattern size, which is projected
onto the wafer, causing a photochemical reaction where it strikes
270 C H A P T E R 8 Nanomaterials: Synthesis and Characterization

High-vacuum chamber the resist. The exposed parts of the resist are dissolved during the
developing stage, as shown in Figure 8.25b, forming a replica of
Collimating Liquid the mask pattern. The assembly is then placed in an acidic solu-
lens metal Ga tion, which attacks the silica but not the resist or the silicon (Figure
ion source
8.25c). Once the silica has been removed, the resist is dissolved in
a different acidic solution (Figure 8.25d). Further etches remove
silicon from the exposed areas, creating channels (Figure 8.25e). The
resulting nanofeatured chip may then be processed further to make
it electronically active or be used as a template for soft lithography,
Focused
described in a moment. Though the concept of photolithography is
Ga ion
beam simple, the implementation is complex and expensive. Masks need
to be perfectly aligned with the pattern on the wafer. The silicon
wafer has to be a perfect single crystal, almost defect-free.
UV light has a wavelength of 250 nm, giving a limiting feature size,
set by diffraction effects, of about 100 nm. Greater resolution is
possible with electron-beam lithography and X-ray lithography. In
FIGURE 8.24
electron-beam lithography, the pattern is written in a polymer film
Focused ion beam machining. The beam is
with a beam of electrons. The shorter wavelength of the electrons
focused to a spot of 5–7 nm diameter, allowing
allows features with a smaller scale than is possible with UV light,
nanoscale shaping.
but the technique is slow and expensive. In X-ray lithography, dif-
fraction effects are minimized by the short wavelength (0.1–10 nm),
but conventional lenses are not capable of focusing X-rays, and the
Collimated UV beam radiation damages many of the materials used for masks.
Because of these limitations, the most recent lithography methods
Mask
make use of mechanical processes—printing, stamping, molding,
Photo-
resist and embossing—instead of photons or electrons. Two of these soft
(a)
Silica, SiO2 lithograph processes are shown in Figure 8.26. The starting point is a
Silicon silicon mold made by photolithography or e-beam lithography, as
in Figure 8.25. Subsequently a chemical precursor to polydimeth-
Dissolve
(b) exposed ylfiloxane (PDMS) is poured over and cured into the rubbery solid
resist PDMS stamp that reproduces the original pattern. The stamp can
Etch then be used in various inexpensive ways to make nanostructures.
(c) silica In the case of microcontact printing the stamp is inked with a
layer
solution consisting of organic molecules and then pressed into a
Remove
(d) remaining thin film of gold on a silicon plate, as shown on the left of Figure
resist 8.26. The organic molecules form a self-assembled monolayer on
Etch the solid surface that reproduces the pattern with a precision of
(e) underlying
silicon approximately 50 nm. In micromolding, the PDMS stamp is placed
on a hard surface and filled with a liquid polymer, as on the right
FIGURE 8.25
in Figure 8.26. The filled mold is then pressed onto the surface of a
Photolithography. A beam of UV light activates the
photoresist, transferring the pattern from the mask
silicon or other wafer; the polymer is polymerized and the pattern
to the sample. is transferred to the wafer with a resolution approaching 10 nm.
The advantage of using soft lithography methods is that, once the
master template has been made, no special equipment is required.
 Characterization of Nanomaterials 271

Soft Soft
(a) PDMS (a) PDMS
template template

(b) (b) Polymer-

Inked inked
template template

Liquid
Gold polymer
layer
(c) (c)
Silicon
Silicon

Printed Solid
surface polymer
(d) (d)

FIGURE 8.26
Two examples of soft lithography processes. In both cases a stamp of PDMS is first fabricated by molding it from a silicon template made in the
way shown in Figure 8.25. Subsequently the stamp can be used for producing nanostructures following the microcontact printing method (left
figure) or the micromolding-in-capillaries method (right figure).

Soft lithographic methods are capable of producing nanostructures

in a wide range of materials and can print or mold on curved as well
as planar surfaces. They are not, however, ideal for nanoelectronics,
because circuits require stacked layers of different materials, and
distortion of soft PDMS stamps can produce errors in the patterns.
The large (and expanding) number of processes that manipulate
materials at the nanoscale are asssembled in the final synthesis
diagrams shown in Figure 8.27 and 8.28. Figure 8.27 summarizes
processes for making nanoclusters, nanolayers and nanofilms, as
well as methods for nanoprofiling to make nanoscale features.
Other processes are under development. Figure 8.28 summarizes
processes that are appropriate for different nanomaterial forms,
including nanoparticles, nanowires and nanotubes, nanofilms, and
bulk forms; and broadly suggests their relative scales. Figure 8.1
shown earlier summarizes processes according to the 0-D, 1-D, 2-D
and 3-D classification system discussed earlier.

8.2 CHARACTERIZATION OF
NANOMATERIALS
In general, the role of characterization techniques is to establish a
correlation between the structure, shape, and chemical composi-
tion of nanomaterials obtained in processing, with their proper-
ties. In the case of nanomaterials, the important aspects to consider
about characterization methods are the type of information and
the resolution achieved by each technique. For each nanomaterial
272 C H A P T E R 8 Nanomaterials: Synthesis and Characterization

Inert gas Vapor phase Sono-chemical Sol-gel

condensation expansion processing deposition

Methods for making

nanoparticles,
Self-assembly - Languimuir nanoclusters and
droplets Blodgett film
nanolayers
- +

M+

Beating Pulsed Physical vapor Chemical vapor

electro-deposition deposition (PVD) deposition (CVD)

Melt spinning Laser surface Electroless nickel Equal angle

hardening process extrusion

Methods for
making bulk
Clusters, nanostructured
particles, chips materials
Powder milling Micro-machining

Consolidation
methods

Pressure sintering Electric discharge

(”Flash” sintering)

Motor

Motor

Motor

Micro-machining Electron beam Femto-second Focused ion beam

workstation machining laser machinning machining
Methods for
nanoprofiling

Photolithography Two soft lithography processes

FIGURE 8.27
Summary of synthesis methods discussed in Section 8.1.
 Characterization of Nanomaterials 273

Inert Gas Condensation

Free Jet Expansion
Sonochemical
Sol-Gel
Particles
Molecular Self-Assembly

Molecular Self-Assembly
C.V.D
Arc Discharge
Wires + Tubes V.L.S

Lithography

Micromachining
Focused Ion Beam Patterning
Electron Beam

Films Thin Film Deposition Electrodeposition

Techniques
Sputtering
Molecular Epitaxy
C.V.D.

Foil Beating
Langmuir Blodget

Bulk Nanomaterials
Bulk Forms
Equiangle Extrusion

Compaction + Sintering

Nanoscale Microscale Macroscale

FIGURE 8.28
Summary of processes used to make nanomaterials in relation to scale.
Excitation
aperture probe
type (0-D, 1-D, 2-D, and 3-D), one or more techniques can be used
to extract the desirable information. A simple way of categorizing Sample
the characterization methods is to consider imaging and analyti-
cal techniques. Imaging involves some kind of microscopy, whereas
analysis involves some type of spectroscopy.
Let’s begin with the imaging techniques. These methods can involve
light, electrons, ions, or scanning probes. With respect to light micro-
scopes, the most common approach for imaging nanomaterials is
the so-called near-field scanning optical microscope (NSOM; see Figure Detection
8.29). In this type of microscopy a subwavelength aperture (smaller
than the wavelength of light) is placed in close proximity to the FIGURE 8.29
sample (∼10 nm). In addition, a light source is used as a scanning Schematic diagram of a near-field scanning optical
probe. To achieve good resolution, the probe is highly focused. The microscope.